Re: FW: Have you seen this? An improved way.

> Now I certainly wish it handled other declaration types, particularly
> numeric types, but I understand why Torronto can't implement that by
> themselves. However, to my knowledge it is disinformation to suggest
> OPDESC does not work correctly when used as designed.  Or at least I've
> never experienced nor do I recall reading on this forum the fact it does
> not work with any combination of CONST or *NoPass or whatever.

It's possible that a PTF has fixed the problems with OPDESC, but my
experience (and this was back on V3R2) was that OPDESC worked MOST but not
all of the time.

I'd have subprocedures in a program that used OPDESC.  They'd work,
getting the length of the variable from the operational desriptor during
testing... then, like 6 months later, that same call (which was made at
the same place in the  same program) would bomb out saying that no
descriptor was passed.   It had worked fine for 6 months, then suddenly
stopped!

Since it's actually LESS code to pass the length as an additional
parameter to the subprocedure, I stopped using OPDESC at that point,
because I can't stand unreliable code.

Since then, I've never used OPDESC -- like I said, it could be that this
bug has been fixed, but in my opinion, Bob is right on the money with the
statement that OPDESC is unreliable.